Hugging Face Connectivity Issues

[zm711]

I wasn't part of the hugging face addition so I'm not sure how important those tests are. we already have to deal with gin and ibl connectivity. Should we do a skip if hugging face is down or do think we need to absolutely check those tests for each PR? Since I didn't follow that specific addition I figured it was better to ask to get clarity for myself :)

related to this test run: https://github.com/SpikeInterface/spikeinterface/actions/runs/14155883866/job/39655087354?pr=3820

[alejoe91]

We can definitely skip, but I'm not sure how to test if HF is down :)

[zm711]

Can we ping from the runners? Maybe I'll play around a bit. There's no rush for this :)

[samuelgarcia]

I would skip and have regular cron ci depending on external platform.

[zm711]

Cool, so we could check for an env variable skip if that isn't there then add that env var to codecov daily... That would work for me if Alessio is cool with it. Hopefully it will speed up the sortingcomponents tests a little since they are taking ~10mins which is ~25% of our testing time.

[samuelgarcia]

About the sortingcomponents, we plan an enormous clean with Pierre in spring.